CDN and cybersecurity giant Cloudflare held its Q2 earnings call on Thursday, during which chief financial officer Thomas Seifert gave a concerning prediction about what the web...
That’s not the point. Let’s put aside that an LLM driven agent is not deterministic ie, not always doing what the user intended, just like regular text autocomplete messes up as well regularly. The point is that like with e-mails when spam became so easy to generate that it flooded the system, now more sophisticated actions on the internet get easier to do (that said, most of it was possible already before, just required more skill), at least as long as LLMs are not charged at real costs but highly subsidised.
At some point something has to give. Either tools will be developed to somewhat reliably classify the nature of the traffic or things will start to be priced out as everyone will have to pay the bill for this new kind of spam (ie low value high volume traffic, or ill intended traffic)
A lot of that was scraping. A huge amount. Just take everything possible. Hopefully that will slow down, or as I did, I allow requests and reponses, but scraping gets kicked.
But it is getting hard and harder to self host. Large websites and hosts (gitlab, self labs) get slammed and even DOS. Its been rough.
There are huge network upgrades in the works to try and deal with this 3.2T are being looked at already. But of course electricity is going to hit a hard ceiling both for the data center and the network center. Hence the costs you are talking about.
I am more concerned with the fact that if everyone is asking AI, who is going to put the answers somewhere to get them? Who is going to self host and publish like I do today? It already is rare to see people running their own websites anymore. Lemmy is pretty amazing for that, but it too has gotta be having a huge amount of traffic at each instance.
I guess I am saying I am aware of your concerns and I have mine too.
It feels a bit like in Cyberpunk, where the wild Internet has become unusable or even dangerous and only big corpo controlled parts are still accessible to the masses.
The other topic you mentioned, LLM’s are going to hit a wall and they’ll do so hard. Training will be increasingly relegated to information predating large scale use if LLMs to avoid rapid degradation of the models by training it with LLM output. That in itself creates a huge problem the more time passes. Nothing is safe btw, analog media are increasingly just as tainted.
That’s not the point. Let’s put aside that an LLM driven agent is not deterministic ie, not always doing what the user intended, just like regular text autocomplete messes up as well regularly. The point is that like with e-mails when spam became so easy to generate that it flooded the system, now more sophisticated actions on the internet get easier to do (that said, most of it was possible already before, just required more skill), at least as long as LLMs are not charged at real costs but highly subsidised.
At some point something has to give. Either tools will be developed to somewhat reliably classify the nature of the traffic or things will start to be priced out as everyone will have to pay the bill for this new kind of spam (ie low value high volume traffic, or ill intended traffic)
A lot of that was scraping. A huge amount. Just take everything possible. Hopefully that will slow down, or as I did, I allow requests and reponses, but scraping gets kicked.
But it is getting hard and harder to self host. Large websites and hosts (gitlab, self labs) get slammed and even DOS. Its been rough.
There are huge network upgrades in the works to try and deal with this 3.2T are being looked at already. But of course electricity is going to hit a hard ceiling both for the data center and the network center. Hence the costs you are talking about.
I am more concerned with the fact that if everyone is asking AI, who is going to put the answers somewhere to get them? Who is going to self host and publish like I do today? It already is rare to see people running their own websites anymore. Lemmy is pretty amazing for that, but it too has gotta be having a huge amount of traffic at each instance.
I guess I am saying I am aware of your concerns and I have mine too.
It feels a bit like in Cyberpunk, where the wild Internet has become unusable or even dangerous and only big corpo controlled parts are still accessible to the masses.
The other topic you mentioned, LLM’s are going to hit a wall and they’ll do so hard. Training will be increasingly relegated to information predating large scale use if LLMs to avoid rapid degradation of the models by training it with LLM output. That in itself creates a huge problem the more time passes. Nothing is safe btw, analog media are increasingly just as tainted.